SUPPORTED BY MULTIPLE PATENTS

AI Agent Security
& Governance

Name: AgentPass
Author: CyberSecAI Ltd

End-to-end compliance for autonomous AI agents. Trust scoring, identity verification, AML screening, and forensic accountability.

IETF: MCPS Draft IETF: Payment Trust Draft AgentSign MCPSaaS Podcast

Screen the Agent

Trust scoring (L0-L4). Is this agent trusted enough to pay?

+ MASTERCARD CHECK

Screen the Recipient

AML screening. 75K+ sanctions entries (UK HMT + OFAC SDN) + Mastercard Onboard Risk Check.

Screen the Jurisdiction

Sanctioned countries blocked. Most-restrictive-applies.

Prove Identity

Challenge-response ECDSA P-256 verification. Per-agent keys.

Forensic Evidence

Hash-chained audit trail. 7-year retention. Court-admissible.

SAR Generation

Automatic suspicious activity reports for financial intelligence units.

Travel Rule

FATF Rec 16 compliance. Originator/beneficiary data for cross-border.

AML Patterns

Detects structuring, layering, smurfing across multiple agents.

Pre-Payment Gate

All checks happen BEFORE the payment touches Stripe or any network.

OPENAPI APPROVED

x-agent-trust extension · live in this demo

This demo signs the payment with ES256 and sends an Agent-Signature header per the OpenAPI extension officially registered at spec.openapis.org/registry/extension/x-agent-trust.html. The server verifies it against the live JWKS at /.well-known/agent-trust-keys · view the spec

Auto-runs on load. Hits live production. Real sanctions data. Real ECDSA signatures. Real x-agent-trust signing. Sandbox account with $10K test balance.

Platform

Docs npm SDK AgentSign MCPSaaS

IETF: MCPS IETF: Payment Trust Podcast

CyberSecAI Ltd (16837162) -- Supported by multiple patents -- IETF Internet-Drafts -- Referenced in OWASP -- Cited in academic research -- FCA Sandbox Submitted

AI Agent Security& Governance

Platform

AI Agent Security
& Governance